![AI. Machine Learning. What’s the Impact on Digital Marketing Today? [PODCAST]](https://cdn.searchenginejournal.com/wp-content/uploads/2017/05/ai-machine-learning-online-marketing-search-engine-nerds-ep132-760x400.jpg)
Low SEO budget? Get all the tools you need – for free!
SEO PowerSuite has helped 2 million users grow their business online, so it will help grow yours. Get your free copy of SEO PowerSuite now!
Podcast: Download
Subscribe: iTunes | Android | RSS
Artificial intelligence and machine learning have been highly predicted trends in marketing and SEO. But how are they changing the industry now? Read on to find out more about the influence of AI on the marketing world in this Search Engine Nerds episode.
Marc Poirier, CEO and co-founder of Acquisio, joins SEJ’s Brent Csutoras to talk about how artificial intelligence and machine learning impacts online marketing.
Poirier also gives us a primer on how AI is affecting local search, and shares his verdict on the battle between AI and humans in search and marketing.
We’re starting to see AI integrate itself into a lot of marketing campaigns. Is that the direction we’re headed in?
Marc: I think this is the direction most software companies need to start steering towards over the next five years or so. We have all this data we’re looking at; no matter what field of marketing you’re in, there’s just so much data coming at us now. Being able to leverage that data and make good use of it through machine learning and data mining is something that’s happening today, but it’s going to accelerate over the next few years.
How do you see the transition of machine learning and AI really start to interact with marketing over the next five years?
Marc: There are various models of algorithms that are designed to solve specific problems, and they exist out there. So you can just use some of the work that has been done by academics before and apply it to the problem you’re trying to solve to parse through a ton of data and come out with a better answer really rapidly. I think a lot of companies do that today, but it hasn’t necessarily been labeled as such.
The reality is not so much about the data and having tools to visualize it. That’s great, but what are you going to do with it? How do you analyze it? How do you make decisions based on that data, and how do you automate that decision making process? How do you do that in marketing?
Brent: I think we see that even with small data. You can go in and show somebody Google Analytics at a very small scale and show the data.
Marc: Yeah, exactly.
Brent: And people aren’t really willing to implement that data as it is, let alone start getting into large-scale data collection and analysis. I think one of the big things for marketers is this sense of disconnect because it might not be overly new. It might not be overly complicated at the rudimentary level, but when it comes down to it, if I wanted to do AI, if I wanted to incorporate some of this stuff into my marketing campaigns, there’s a disconnect in the path on how I do that.
When you look at an AI solution versus an expert, do you feel that AI provides an equal or better outcome? Or is it still just a time versus money type of scenario?
Marc: No, it’s a better outcome, but it depends on the problem you’re solving. Some of the things we do are just research, so the outcomes are not better.
What we’ve been doing for five or six years is controlling money, and that’s not something humans can do very well, especially at scale. So we build software to help the small business, but we don’t sell it directly to them. We work with large resellers, companies that have tens of thousands of small businesses they need to serve. They need to make sure that the results are on point, that they’re spending the budgets, and they’re maximizing phone calls and clicks and things like that. This is where technology will do a better job than humans 100% of the time.
Brent: Is this similar to the stock market where you really have no chance if you’re not utilizing these tools? Has it gotten to that point, you feel?
Marc: Yeah, it has. We’re making decisions and making changes on budgets and bids every 30 minutes now across tens of millions of ad groups and hundreds of thousands of accounts. So you can’t imagine humans doing that.
First of all, it’s redundant. It’s error-prone and there’s a lot of accurate decision making going on, a lot of machine learning applied that always tries to hone in on the target. We’re trying to spend a $500 budget, or a $200 budget — it’s very little money, so how do we hit it? How do we get really close to it and not go over? How do we learn about that specific business, about their trend? So there are a few levers there that are really important to model.
I mean you could do it if you have just one account to manage and you apply a lot of time and energy. I’m sure humans can do a pretty good job at it, but you can’t do that at scale for small accounts. It just doesn’t make any sense, and it doesn’t work. It’s not possible to make the right decision all the time. That’s what algorithms and AI do.
What would be the steps and tools to start incorporating AI into paid search? What would be the end-point solution for somebody right now?
Marc: We have a product called Promote where any local business can sign up and we’ll create a local presence for them. We’ll create a Google AdWords account for them and all of the keywords they need, the ads, everything. It’s all automated. Then we’ll assign a small budget and control it really well.
If you look at what Google’s doing, or Marin Software, Kenshoo, Adobe, Ignition One, SearchForce, all the companies in this space, all the bid management solutions have some kind of algorithms or rule engines. So there is technology there to help make the right decisions on large data sets, and they do very well for different situations.
There are products out there in the search world that do that, but they’re limited, typically, to bid management. I think the areas of focus for us are how do we go beyond that in the future and automate more and do a better job at writing ad copy, for example, or testing ad copy.
What is it about AI or machine learning that’s really impacting local search? How does it affect what people are doing locally?
Marc: It revolves a lot around delivering value for small budgets, so it’s very difficult to control. If you have a lot of money to spend on AdWords, you’ll be able to see results, and there’s going to be a lot of waste. But you can’t afford any waste when you have a $200 budget for the month, right? You want to get some clicks. You want to get some calls, so it’s really important that we go out there and get those clicks that will cost the least amount of money. They need to be valid clicks and they need to be relevant.
It’s understanding the auction, understanding the profile for each business — the peaks and valleys, time of day, day of week — to see how the data has flowed historically when we have that. This is something the software will pick up, not only per vertical but also by business. You just gain more certainty if you can have many layers.
So you can say, “Okay, I’m looking at the law services category in general, all of the different attorneys and legal services that are available. Do I see patterns on a day of week basis? Yes. I see some kind of trend where Monday seems to be maybe 20% busier than all the other days.”
Then let’s look at by city for example, or by DMA. Do we see trends there as well? Yeah, larger cities seem to have more of that. Then per business, like for this specific business, if I look back three months, do I see a pattern? Do all these things come together? And does that help me make a better decision? How sure am I that this decision of changing that bid, lowering it by five cents will have an impact in the next 30 minutes?
In a nutshell, it’s a lot of complexities, a lot of data points to look at, but it’s to make that small decision. They’re always small changes that we make, and then validate the next 30 minutes, 60 minutes, 90 minutes. Did I see pick up on impressions? Because when there’s not a lot of click-flow, do I at least see more impression flow hour per hour? Those decisions will be tweaked as the function of the results we observe.
How do you feel about where people spend their money on paid ads right now? Do you feel that Google AdWords is still a major place to spend budget? Are you seeing certain social areas better than others?
Marc: Facebook ads is the place where small businesses are going to start because Facebook is making it easy for anyone to start and they’re making it part of your everyday experience. If you have a business page and you wrote a post, we’ll suggest you boost it for 10 bucks or something. Facebook’s been really good at garnering small span from small advertisers, small SMBs. Most SMBs today have realized that before they graduate to do some paid search investments, they’ll go to Facebook first then introduce Google AdWords.
Brent: Is that just because of the ease of use? It’s just something they can easily engage with, and then once they see some kind of traction, they start looking at other areas to run ads?
Marc: Well, yes, it’s easy and there are many ad products on Facebook. It’s a Swiss army knife of advertising. And there are different ways to build your audience as well, so it’s not only the ad products, it’s how you target. It’s quite complex, despite how easy they make it for a small advertiser to start. They have products like local awareness ads that are that simple. But local awareness helps retailers, restaurants and bars, the food industry in general.
But maybe it’s not that helpful for an emergency plumbing service or the landscaper. When you need a roofer, you’re going to look for it. It’s not going to be an awareness play so it may not be the first place you go for those. Depending on the type of business you are, it may be best to start on Facebook when you’re building local awareness. When it’s something that’s more search-driven intent, search ads would be more appropriate to start with.
Do you feel that you can take learnings from one platform to the other? Or do you feel that they really are individual and need to be viewed as individual?
Marc: We’re not finding any learnings from Facebook can apply to search or vice versa. However, the one thing we’re doing there in terms of bringing things together is budget allocation. How do we devise or change the way our product works to automatically allocate the money correctly across search — AdWords, Bing Ads, Facebook? We didn’t crack that nut yet. This is something that’s really interesting to a lot of companies.
But we’re in the middle of exploring if it does make sense, if it adds value. We’re hearing a lot of “yes”. So we’ll continue that investigation, but the idea of having a marketing budget for advertising online, like a digital advertising budget, and to have one place where you can just sort of put your money in there and trust that the algorithm will do the right thing, seems to be appealing to marketers today.
To listen to this Search Engine Nerds Podcast with Marc Poirier:
- Listen to the full episode at the top of this post
- Subscribe via iTunes
- Sign up on IFTTT to receive an email whenever the Search Engine Nerds podcast RSS feed has a new episode
- Listen on Stitcher, Overcast, or Pocket Casts
Think you have what it takes to be a Search Engine Nerd? If so, message Loren Baker on Twitter, or email him at loren [at] searchenginejournal.com. You can also email Brent Csutoras at brent [at] alphabrandmedia.com.
Visit our Search Engine Nerds archive to listen to other Search Engine Nerds podcasts!
[“Source-ndtv”]
